Figure 9From: Performance Analysis of Bit-Width Reduced Floating-Point Arithmetic Units in FPGAs: A Case Study of Neural Network-Based Face Detector Comparison between analytical output errors and experimental output errors. Back to article page